The role of living roofs in holistic stormwater management systems - Living Roofs in Integrated Urban Water Systems - page 28

2.5.1 Hydrologic studies at building scale
Field monitoring of full and pilot scale extensive living roofs show a substantial
contribution to site runoff volume control, with studies across the United States,
Germany, Belgium, Sweden and New Zealand reporting approximately 50-80
percent of precipitation retained over extended periods of data collection (Bergh-
age et al. 2007; Carson et al. 2013; Carter and Rasmussen 2006; Fassman-Beck
et al. 2013a; Getter et al. 2007; Gregoire and Clausen 2011; Hathaway et al.
2008; Mentens et al. 2006; Monterusso et al. 2004; Stovin et al. 2012; Villarreal
and Bengtsson 2005). A notable exception was measured at 36 percent retention
from a system in New York City primarily comprised of synthetic water retention
mats and only 32 mm of growing media (Carson et al. 2013). Interpreting percent
retention statistics must account for the duration of the study, or more specii-
cally, rainfall characteristics, as the mathematics of calculating a percent differ-
ence can create bias. Stovin et al. (2012) comment that because of the frequency
with which runoff is fully retained, percent difference statistics of either retention
or peak low reduction are somewhat meaningless. Many small events arbitrarily
elevate statistics, and vice versa. Time of year of the study is important as periods
with relatively higher or lower rainfall and ET will also be relected in individual
event performance ( Section 2.6 discusses ET in detail) (Berretta et al. 2014;
Schroll et al. 2011; Wadzuk 2013). That being said, the importance of seasonal
performance variability depends on interpretation. In Auckland, New Zealand,
winter event retention was maintained at 66 percent, which is still considered
signiicant in terms of stormwater management (Fassman-Beck et al. 2013a).
In general, excellent performance for small storm events is usually observed, with
reduced retention in very large, infrequent events (Carpenter and Kaluvakolanu
2011; Carson et al. 2013; Carter and Rasmussen 2006; Fassman-Beck et al. 2013a;
Stovin et al. 2012; Teemusk and Mander 2007). High performance variability is typi-
cally reported on an event-to-event basis, with most studies ranging from less than
10 percent or up to 100 percent retention for individual events. Many studies iden-
tify a minimum threshold event smaller than which runoff was never observed,
regardless of climate conditions. Due to the infrequency of occurrence of storms
larger than ~25 mm, few studies have thoroughly characterized performance for
conventional design storm magnitudes. A living roof test plot in the UK retained 50
percent of the annual rainfall, but 30 percent of rainfall for storms larger than the
one-year ARI (Stovin et al. 2012). Results are relatively consistent with performance
of four pilot and full-scale living roofs in Auckland, whose average per-event reten-
tion eficiency decreased by about 20-40 percent for storms greater than 25 mm
(but showed signiicant variation amongst sites) (Fassman-Beck et al. 2013a). From
three full-scale living roofs in New York City, a combination of seasonal retention
performance and storm size was inluential for events 10-40 mm, but not for events
with greater or less rain. It is noted that only about 5-10 percent of observed events
were greater than 40 mm (Carson et al. 2013).
As with rainfall retention, a wide range of peak low reduction eficiencies
have been reported, likely in response to variation in climate, conditions on the